Getting started on NVwulf

Log in

Connect by SSH (DUO two-factor required — see Connecting to the cluster for Windows clients and DUO details):

ssh -X NetID@login.nvwulf.stonybrook.edu

Or use the web portal — Open OnDemand for NVwulf — with your NetID and password.

Transfer data from SeaWulf

Use rsync, run from NVwulf. A single file:

rsync -v NetID@login.seawulf.stonybrook.edu:~/test.txt ~/test.txt

A directory (including subdirectories):

rsync -av NetID@login.seawulf.stonybrook.edu:~/test/ ~/test/

Load software modules

# list what's available
module avail
# load a module
module load miniconda/3

Submit your first GPU job

Load Slurm and check the partitions (h200x4, h200x8, b40x4, plus -long and debug- variants):

module load slurm
sinfo

An example script requesting 1 GPU, 8 CPUs, and 25 GB of memory:

#!/bin/bash
#SBATCH --job-name=test-tf
#SBATCH --output=res.txt
#SBATCH --ntasks=8
#SBATCH --cpus-per-task=1
#SBATCH --nodes=1
#SBATCH --time=05:00
#SBATCH -p h200x4
#SBATCH --mem=25g
#SBATCH --gpus=1

module load tensorflow/2.19.0
python tf_test_nn_training.py

Submit and monitor:

sbatch test_job.slurm
squeue -u NetID